Hotspots_COVID Supplement II

摘要 这项应用是为了应对迫切需要了解的流行病学影响 美国新批准的儿童SARS-CoV-2疫苗接种率不统一。尽管 由于新冠肺炎的持续感染和住院，接种疫苗毫无疑问 带来了显著改善的结果。随着最近批准为年轻人接种疫苗 孩子们，我们现在有了一个真正的机会来控制新冠肺炎。然而，有几个因素可能 为实现这一目标设置障碍：反对疫苗和其他政策，如口罩， 入住率限制和测试(特别是在学校)、新的变种和免疫力减弱。结果, 在接下来的几个月里，即使是接种良好疫苗的亚群，如老年人，风险也可能会增加。 对这种增加的风险进行建模，并找到最脆弱的子组是一项非常具有挑战性的工作 对公众健康有很大影响的问题。 我们的项目将开发一种严格的数据和模型驱动的方法来识别和表征 最脆弱的亚群，并研究降低其风险的战略；这一方法将是 为弗吉尼亚州进行了评估。团队成员一直在支持弗吉尼亚州卫生部(VDH) 新冠肺炎自大流行开始以来的反应，并拥有新的高分辨率数据集和 通过此协作获得领域专业知识。 我们将扩展一个详细的，年龄分层的基于病原体的模型(ABM)，用于SARS-CoV-2的传播。 通过整合关于疫苗迟疑、免疫接种和学校的详细数据集，Sion在弗吉尼亚州 混合和干预。这种反辐射导弹的标定、仿真和分析是一个很困难的问题 由于它的规模很大，这本身就是一个很大的问题。使用这样的模型来确定在这样的 大型模型在计算上非常具有挑战性，这是以前从未做过的。创新的AS- 该项目的目标是将人工智能和高性能计算的工具结合在一起， 针对这类问题扩展ABM。该项目将在很大程度上建立在开发的工具和模型上 由团队成员作为家长资助的一部分。

Summary This application is in response to the urgent need to understand the epidemiological impact of non-uniform uptake of the newly approved SARS-CoV-2 vaccine for children in the US. Despite the continuing infections and hospitalizations due to COVID-19, there is no doubt that vaccination has led to significantly improved outcomes. With the recent approval of vaccinations for young children, we now have a real opportunity for controlling COVID-19. However, several factors could pose roadblocks towards this goal: opposition to the vaccine and other policies such as masking, occupancy limits and testing (especially in schools), new variants and waning immunity. As a result, the risk for even well vaccinated subgroups, such as seniors, could increase in the coming months. Modeling this increased risk, and finding the most vulnerable subgroups is a very challenging problem with high public health impact. Our project will develop a rigorous data and model driven approach to identify and characterize the most vulnerable subgroups, and study strategies for mitigating their risk; this approach will be evaluated for Virginia. Team members have been supporting the Virginia Dept of Health’s (VDH) COVID-19 response since the start of the pandemic, and have novel high resolution datasets and domain expertise through this collaboration. We will extend a detailed, age-stratified agent based model (ABM) for SARS-CoV-2 transmis- sion in Virginia by incorporating detailed datasets on vaccine hesitancy, immunization, and school mixing and interventions. The calibration, simulation and analysis of such an ABM is difficult in itself due to its large scale. Using such a model to identify the most vulnerable populations in such large models is computationally very challenging, and has not been done before. An innovative as- pect of the project is the combination of tools from AI and high performance computing with large scale ABMs for such problems. The project will build significantly on tools and models developed by team members as part of the parent grant.